One of the reason I am skeptical about React is the end quality of visual appearance and user interaction of the React apps. It feels like most of such apps’ developments were led by developers themselves and not by UI/UX people (or by poorly educated or not experienced UI/UX people then).
Even take for example MaterialUI or BootstrapUI (as most popular UI frameworks among React users) which not only both look ugly (yes, I know I am too opinionated here, but have to say it clear and loud) but also very often don’t really provide some basic things (like enhanced dropdown for example).
Semantic UI is not ideal but it is the most minimalistic (in terms of UI itself, not development) framework. Yes, it comes with jQuery and a piece of custom js code for some manipulation, like forms validation or visibility (which both are super-cool out of the box solutions). But this approach seem to contradict with React philosophy of ‘naked components’.
In this regard Semantic fits nicely into Blaze development - we have a nice callback ‘onRendered’ per template, where we can do whatever we want with whatever existing js libraries (and there are plenty of them). Yes, this approach maybe is not rendering as fast or is not as ‘clean’ under the hood as React might be, but c’mon at least we get it done fast and end result is far superior to all the custom quickly developed UI React junk that is out there in the web…
Long story short - if you integrate Semantic UI, jQuery, a pair of ‘heavy’ js libraries (like grids, trees or charts for example) to your React app, what kind of advantage React is going to give you after all if most of user interaction will be done using the plain DOM manipulation js libraries?